Development of a novel microphysiological system for neurotoxicity prediction using human iPSC-derived neurons with morphological deep learning

Abstract

Microphysiological system (MPS) is an in vitro culture technology that reproduces the physiological microenvironment and functionality of humans, and is expected to be applied for drug screening. In this study, we developed a MPS for structured culture of human iPSC-derived neurons, then predict drug-induced neurotoxicity by morphological deep learning. For human iPSC-derived cortical neurons, after administration of three different amyloid {beta} (A{beta}) peptides, neurotoxicity effects were evaluated by a deep learning image analysis, training two artificial intelligence (AI) models on neurites and PSD-95 images. The combined results indicated that A{beta} 1-42 and 1-40, but not 1-28, induced synaptic degeneration, which is close to clinical reports. For human iPSC-derived sensory neurons, after administration of representative CIPN-related anti-cancer drugs, the toxic effects on soma and axons were evaluated by AI model using neurites images. Significant toxicity was detected in positive drugs and could be classified by different effects on soma or axon, suggesting that the current method provides an effective evaluation of chemotherapy-induced peripheral neuropathy. Taken together, it suggests that the present MPS combined with morphological deep learning is a useful platform for in vitro neurotoxicity assessment.